Abstract
Paroxysmal epileptiform abnormalities on EEG are the hallmark of epilepsies, but it is uncertain to what extent epilepsy and background EEG oscillations share neurobiological underpinnings. Confounding factors, including the heterogeneous etiology of epilepsies and medication effects hamper studies on background brain activity in people with epilepsy. This limitation may be partly overcome by assessing the genetic overlap between epilepsy and EEG power spectra based on genome-wide association studies (GWASs). We used the summary statistics of the recently published EEG (n=8,425) and epilepsy (n=37,021) GWASs to compute estimates of genetic overlap between EEG power spectra and epilepsy subtypes using linkage disequilibrium score regression (LDSC) and polygenic risk scoring (PRS). We found significant genetic correlations between genetic generalized epilepsy (GGE) and EEG beta power (rg=0.44; p=0.01) and between GGE and theta power (rg =0.25; p=0.02), which were replicated in an independent GGE cohort. Supporting this finding, beta and theta powers’ PRSs were both strongly and positively associated with generalized epilepsy (p-values ranging from 5.3×10−8 to 8.5×10−4), suggesting that genetic predisposition to GGE and to altered background brain activity partially overlap. All PRS directions of effect were replicated in an independent cohort of people with epilepsy and controls (n=25,303; p=0.023). Furthermore, subjects in the highest 10% quantile of beta-power PRS scores were 1.4 fold more likely to suffer from GGE compared to subjects in the lowest 10% quantile (OR: 1.40; 95%CI: 1.18-1.67; p=1.5×10−4). Moreover, by meta-analyzing EEG and epilepsy GWASs data for gene-based tests and by imputing gene expression, we found increased enrichment of brain-expressed genes relative to both the epilepsy and the EEG GWASs individually. Finally, Mendelian Randomization analyses suggested a causal effect of GGE genetic liability on beta oscillations. In conclusion, our results point to shared biological mechanisms underlying background EEG and the susceptibility for seizures. Our findings thus open avenues to investigate the clinical utility of theta and beta oscillations in GGE. They may also inspire preclinical research into the links between background EEG oscillations and epilepsies.